ner_twitter_fine_tune

This model is a fine-tuned version of distilbert/distilbert-base-cased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4211
• Precision: 0.6078
• Recall: 0.5901
• F1: 0.5988
• Accuracy: 0.9308

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	56	0.3586	0.6358	0.5660	0.5989	0.9323
No log	2.0	112	0.3618	0.6069	0.5746	0.5903	0.9297
No log	3.0	168	0.3722	0.5956	0.6038	0.5997	0.9306
No log	4.0	224	0.3993	0.6060	0.5883	0.5970	0.9301
No log	5.0	280	0.4102	0.5411	0.6329	0.5834	0.9232
No log	6.0	336	0.4077	0.6097	0.5815	0.5953	0.9319
No log	7.0	392	0.4096	0.5858	0.6089	0.5971	0.9286
No log	8.0	448	0.4169	0.5975	0.5832	0.5903	0.9297
0.0111	9.0	504	0.4208	0.6064	0.5866	0.5963	0.9309
0.0111	10.0	560	0.4211	0.6078	0.5901	0.5988	0.9308

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.0
• Tokenizers 0.15.2